Evaluating probability threshold k-nearest-neighbor queries over uncertain data

In emerging applications such as location-based services, sensor monitoring and biological management systems, the values of the database items are naturally imprecise. For these uncertain databases, an important query is the Probabilistic k-Nearest-Neighbor Query (k-PNN), which computes the probabilities of sets of k objects for being the closest to a given query point. The evaluation of this query can be both computationally- and I/O- expensive, since there is an exponentially large number of k object-sets, and numerical integration is required. Often a user may not be concerned about the exact probability values. For example, he may only need answers that have sufficiently high confidence. We thus propose the Probabilistic Threshold k-Nearest-Neighbor Query (T-k-PNN), which returns sets of k objects that satisfy the query with probabilities higher than some threshold T. Three steps are proposed to handle this query efficiently. In the first stage, objects that cannot constitute...

• Type: Conference
• Year: 2009
• Where: EDBT
• Authors: Reynold Cheng, Lei Chen 0002, Jinchuan Chen, Xike Xie
• Keywords: Database, EDBT 2009, K-nearest-neighbor Query, Probabilistic K-nearest-neighbor Query, Threshold K-nearest-neighbor Query